The JOBS Act and Information Uncertainty in IPO Firms

The JOBS Act and Information Uncertainty in IPO Firms
Mary E. Barth*
Graduate School of Business
Stanford University
Wayne R. Landsman
Kenan-Flagler Business School
University of North Carolina
Daniel J. Taylor
The Wharton School
University of Pennsylvania
July 2014
___________________________
*Corresponding author: Graduate School of Business, Stanford University, 94305-5015,
[email protected]. We thank Ernst & Young LLP (EY) for providing data on Emerging
Growth Companies and their disclosures; the views expressed in this paper are the views of the
authors and do not represent in any way the views of EY. We also thank Vikram Dhawan and
David Wu for research assistance. We appreciate funding from the Center for Finance and
Accounting Research, Kenan-Flagler Business School, the Center for Global Business and the
Economy, Stanford Graduate School of Business, and The Wharton School.
The JOBS Act and Information Uncertainty in IPO Firms
Abstract
This study examines whether the Jumpstart Our Business Startups Act (JOBS Act) increases
information uncertainty in IPO firms. The JOBS Act creates a new category of issuer, the
Emerging Growth Company (EGC), and eases disclosure requirements for IPO firms with EGC
status. Measuring information uncertainty using IPO underpricing and post-IPO return volatility,
we find that both underpricing and volatility are significantly greater for IPO firms with EGC
status than comparable firms without EGC status. Additional findings indicate that variation in
the application of specific provisions of the JOBS Act explains the differences in underpricing
and volatility. Taken together, the findings indicate that the JOBS Act’s eased disclosure
requirements increased information uncertainty in IPO firms.
JEL Classification: D8, G14, G18, G32, M41
Keywords:Initial public offering; Information uncertainty; JOBS Act; IPO underpricing; PostIPO equity volatility
The JOBS Act and Information Uncertainty in IPOs
1. Introduction
The question this study addresses is whether the Jumpstart Our Business Startups Act
(JOBS Act) increases information uncertainty in firms with initial public offerings (IPOs). The
JOBS Act, which was signed into law in April 2012, creates a new category of issuer, the
Emerging Growth Company (EGC), and eases regulations for EGCs to encourage initial public
offerings of their shares. Specifically, the Act includes provisions that allow firms with EGC
status to reduce the scope of mandatory disclosure of financial statement and executive
compensation information, to file draft registration statements confidentially with the Securities
and Exchange Commission (SEC), to delay application of new or revised accounting standards,
and to delay compliance with Section 404(b) of the Sarbanes-Oxley Act (SOX), which relates to
auditor attestation on internal controls. We find evidence consistent with the easing of these
regulations increasing information uncertainty in the IPO market.
A key purpose of the JOBS Act is to reduce the cost of accessing capital markets by
eliminating unnecessary or overly burdensome regulations, thereby reducing the cost of an IPO.
In reference to the JOBS Act, the SEC’s website states that “…a cost-effective access to capital
for companies of all sizes plays a critical role in our national economy.” The JOBS Act reflects
the belief that enabling more companies to gain less costly access to capital would enable more
companies to create jobs (hence the acronym) throughout the economy. A March 23, 2012
editorial in the Washington Post expresses this logic succinctly: “The case for the JOBS Act goes
like this: Small companies create jobs. The easier it is to fund a small company, the more jobs
there will be. Federal rules make it harder for start-ups to raise capital. Ergo, relax the rules.”
However, critics of the JOBS Act warned that easing regulatory requirements could
undermine investor protections, thereby exposing investors to greater information uncertainty in
IPO firms.1 This claim is consistent with extant theory suggesting that if mandatory disclosure
provides investors with decision-relevant information, reducing the disclosure will increase
information uncertainty for investors. One consequence of an increase in uncertainty is an
increase in the risk premium that investors demand to compensate them for holding the firms’
shares. Thus, even though the JOBS Act may decrease compliance costs associated with an IPO,
the Act could increase the costs of an IPO by increasing the risk premium, and therefore the IPO
firm’s cost of capital. Examining whether the JOBS Act is associated with an increase in
information uncertainty for IPO firms with EGC status provides insight as to whether the critics’
claims have merit.
To qualify for EGC status, an IPO firm must have less than $1 billion in annual revenue
in the year prior to its IPO. This revenue threshold is sufficiently high that there are few firms
without EGC status after the JOBS Act. Therefore, to address our research question, we
compare information uncertainty in IPO firms with EGC status to that in IPO firms that would
have qualified for EGC status had their IPO occurred after the JOBS Act, i.e., firms that meet the
$1 billion revenue threshold but had IPOs before the JOBS Act. Because all firms in our tests
meet the JOBS Act’s $1 billion revenue threshold, selection based on that threshold cannot
account for any differences in information uncertainty we observe between the two groups of
firms.
1
In a March 14, 2012 letter to the Senate Banking Committee, SEC Chairman Mary Shapiro warned: “Too often,
investors are the target of fraudulent schemes disguised as investment opportunities…if the balance is tipped to the
point where investors are not confident that there are appropriate protections, investors will lose confidence in our
markets, and capital formation will ultimately be made more difficult and expensive.” 2
We test for differences in information uncertainty using linear regression and propensity
score matching. Following prior research, we measure information uncertainty using IPO
underpricing and post-IPO equity return volatility. We employ three measures of underpricing:
market-adjusted stock returns from the offer price to the closing price on the day of the IPO,
from the offer price to the closing price on the day after the IPO, and from the offer price to the
closing price thirty days after the IPO. We find that all three measures of underpricing are larger
for IPO firms with EGC status. We employ three measures of post-IPO volatility: total
volatility, idiosyncratic volatility, and beta, all of which are estimated over the thirty days after
the IPO. We find that total volatility and idiosyncratic volatility are significantly higher for IPO
firms with EGC status, but no evidence of a difference in beta. These findings are consistent
with IPO firms with EGC status having greater information uncertainty at the time of their IPOs
and that the greater information uncertainty pertains primarily to idiosyncratic information.
We use two approaches to test for whether variation in the extent to which IPO firms with
EGC status apply provisions of the JOBS Act explains the larger IPO underpricing and higher
post-IPO volatility. There is substantial variation in the number of IPO firms with EGC status
that elect to apply each provision. For example, whereas 87% omit a compensation discussion
and analysis, only 43% present fewer than three years of audited financial statements. Finding
that specific provisions of the Act are associated with differences in IPO underpricing and postIPO volatility between the two groups of firms increases our confidence that the differences are
attributable to the JOBS Act.
In our first approach, we construct an index based on the number of provisions that each
IPO firm with EGC status applies. We find that the index explains the larger underpricing and
higher total and idiosyncratic volatility for IPO firms with EGC status, and that simply having an
3
IPO after the JOBS Act does not. In our second approach, we examine the relation between
application of specific provisions of the Act and underpricing and volatility. We find that several
individual provisions, most notably confidentially filing draft registration statements, presenting
compensation information for fewer than five top executives, and presenting fewer than three
years of audited financial statements incrementally explains the larger IPO underpricing and
higher volatility for IPO firms with EGC status. Collectively, these findings increase our
confidence that the greater information uncertainty in IPO firms with EGC status is attributable
to the JOBS Act.
The remainder of the paper is organized as follows. Section two discusses the
institutional setting and provides the basis for our predictions. Section three explains the
research design, section four describes the sample and data, and section five presents the results.
Section six offers a summary and provides concluding remarks.
2. Institutional Setting and Basis for Predictions
2.1 JOBS Act
The Jumpstart Our Business Startups Act (JOBS Act) was signed into law on April 5,
2012. The JOBS Act contains seven titles. The portion of the JOBS Act that is relevant to our
study is Title I, which creates a new category of issuer, the Emerging Growth Company (EGC).2
To qualify for EGC status, an IPO firm must have total gross annual revenues of less than $1
billion during the most recent fiscal year ending prior to the IPO. EGC status is maintained until
the earliest of: (i) the last day of the fiscal year in which annual gross revenues are $1 billion or
2
Title II lifts the prohibition against solicitation or advertising of new issues to “qualified buyers” (see SEC Release
No. 33-9415, Final Rules). Titles III and IV direct the SEC to issue new rules exempting “crowdfunding” offerings
of up to $1 million, and securities offerings of up to $50 million, from SEC registration requirements (see SEC
Release Nos. 33-9470 and 33-9497, Proposed Rules). Titles V and VI raise the shareholder registration threshold
for public company reporting from 500 shareholders of record to 2,000 shareholders of record, and Title VII directs
the SEC to conduct outreach to small and minority owned business to inform them of the changes made by the Act.
4
more, (ii) the last day of the issuer’s fiscal year following the fifth anniversary of its IPO, (iii) the
firm issues more than $1 billion in non-convertible debt over the prior three-year period, (iv) the
firm becomes a large accelerated filer, i.e., has public float in excess of $700 million. Once it is
lost, EGC status cannot be regained.
Title I includes various provisions that ease IPO disclosure requirements for firms with
EGC status. IPO firms with EGC status can elect to apply some or all of these provisions. First,
the JOBS Act allows firms with EGC status to file draft IPO registration statements
confidentially with the SEC, provided that the filing and any amendments to it are filed publicly
no later than 21 days before the firm conducts a road show. Before the JOBS Act, IPO firms
could not confidentially file their draft registration statements.3
Second, the JOBS Act provides for a reduction in disclosure of executive compensation
and the number of years of audited financial statements in the IPO registration statement. In
particular, prior to the JOBS Act, IPO firms were required to disclose three years of
compensation information for the Named Executive Officers, i.e., the CEO, CFO, and the three
other highest paid executives, and to provide a compensation discussion and analysis. After the
JOBS Act, IPO firms with EGC status are required to disclose compensation information only
for two years and only for three named executives, including the CEO, and are not required to
present a compensation discussion and analysis. Additionally, before the JOBS Act, IPO firms
were required to include in their registration statements audited financial statements for three
years or for the life of the company, if shorter. After the JOBS Act, IPO firms with EGC status
are required to include only two years of audited financial statements.
Third, the JOBS Act allows firms with EGC status to delay application of some
accounting standards and delay compliance with Section 404(b) of SOX. Before the JOBS Act,
3
Some privately listed foreign firms and government-owned foreign firms were exempt from this restriction.
5
IPO firms were required to apply new or revised accounting standards at their effective dates for
public companies. After the JOBS Act, IPO firms with EGC status are not required to comply
with any new or revised financial accounting standard until it applies to non-public companies.4
Regarding SOX compliance, before the JOBS Act, IPO firms were required to comply with
Section 404(b) of SOX that mandates auditor attestation of the effectiveness of internal controls
over financial reporting beginning with the second annual report after the IPO.5 After the JOBS
Act, IPO firms with EGC status are permitted to opt out of compliance with Section 404(b) of
SOX for up to five years.
Additionally, the JOBS Act exempts IPO firms with EGC status from future auditing
standards adopted by the Public Company Accounting Oversight Board (PCAOB), unless the
SEC decides that such standards should apply to EGCs; advisory say-on-pay votes required by
the Dodd-Frank Wall Street Reform and Consumer Protection Act of 2010; and quiet period
restrictions on research coverage by affiliated analysts. The JOBS Act also provides
underwriters of IPO firms with EGC status greater freedom to communicate with qualified
buyers as referred to in the JOBS Act to determine their interest in the offering.
2.2 Related Literature and Empirical Predictions
There is an extensive literature regarding IPOs, including the pricing of IPO shares and
post-IPO performance (see Ritter and Welch, 2002 for a review). This literature establishes that
the average IPO is underpriced, i.e., the closing price on the day of the IPO is greater than the
4
That is, firms with EGC status are permitted to apply new or revised accounting standards issued by the Financial
Accounting Standards Board as if they were private companies. Although firms with EGC status may irrevocably
elect to comply with these standards when the standards are effective for public companies, they are not permitted to
apply the standards selectively, i.e., standard by standard.
5
Non-accelerated filers, i.e., firms with public float less than $75 million, were exempt from compliance with
Section 404(b) of SOX even after two years.
6
IPO offer price, and that the underpricing is substantial.6 IPO underpricing represents a cost of
equity capital because it reflects “money left on the table” during the offering, i.e., the amount of
money that accrues to the investors in the firm’s IPO to compensate them for holding the shares.
Thus, underpricing represents a risk premium that investors demand to compensate them for their
uncertainty about the value of the firm (Beatty and Ritter, 1986; Rock, 1986). Beatty and Ritter
(1986) and Rock (1986) show analytically that as investors’ uncertainty about the value of the
firm increases, the required risk premium increases, which in turn increases the amount of IPO
underpricing. The subsequent analytical and empirical IPO literature supports the interpretation
of underpricing as a risk premium (see Ljundqvist, 2007 for a review). For example, Lowry,
Officer, and Schwert (2010) uses post-IPO equity return volatility as a measure of uncertainty
and finds that firms with larger underpricing have higher post-IPO volatility.
Prior research finds that increasing voluntary disclosure can reduce IPO underpricing,
which is consistent with the underpricing being a risk premium for uncertainty. For example,
Beatty (1989) and Willenborg (1999) show that higher quality auditors can reduce underpricing
by increasing the quality of the financial statements in the IPO prospectus. Schrand and
Verrecchia (2005) finds that IPO firms with more frequent disclosures prior to the IPO have
smaller underpricing. Leone, Rock, and Willenborg (2007) finds that IPO firms disclosing more
information about the use of proceeds have smaller underpricing. Arnold, Fishe, and North
(2010) finds that firms with a greater proportion of “hard” relative to “soft” information in their
IPO registration statements have smaller underpricing. Boulton, Smart, and Zutter (2011) finds
that IPO firms in countries with higher earnings quality have smaller underpricing.
6
For example, Ritter and Welch (2002) reports an average first-day return of 18.8 percent for a comprehensive
sample of IPOs between 1980 and 2001. 7
Theory predicts that if mandatory disclosure is informative, a reduction in mandatory
disclosure increases uncertainty among investors (e.g., Lambert, Leuz, and Verrecchia, 2007).
Because the JOBS Act has several provisions that ease disclosure requirements for IPO firms
with EGC status, we expect these firms have greater information uncertainty than firms without
EGC status. Taken together, the related literature leads us to predict that this greater information
uncertainty manifests as larger IPO underpricing and higher post-IPO return volatility for firms
with EGC status.
3. Research Design
3.1 Overview
A key feature of the JOBS Act that directly affects our research design is that to qualify
for EGC status, and therefore to be eligible to apply the eased regulations in the Act, an IPO firm
must have less than $1 billion in annual revenue in the year prior to its IPO. As section 4
explains, the $1 billion revenue threshold is sufficiently high that the vast majority of firms with
IPOs after the JOBS Act qualify for EGC status, virtually all of which elect to apply some
provision of the JOBS Act. Thus, there are few IPO firms without EGC status after the JOBS
Act.7
Therefore, to address our research question, we compare information uncertainty in IPO
firms with EGC status to that in IPO firms that would have qualified for EGC status had their
IPO occurred after the JOBS Act, i.e., firms that meet the $1 billion revenue threshold but had
IPOs before April 2012. For example, a firm with an IPO in November 2011 with less than $1
billion in revenue would have qualified as an EGC if it had an IPO seven months later in June
7
See also Ernst &Young LLP (2013a, 2013b) and Latham and Watkins (2013, 2014). 8
2012.8 We refer to firms with EGC status as “EGC firms”, and firms that would have qualified
for EGC status had their IPO occurred after the JOBS Act as “NEGC firms.”
We use two approaches to test for differences in information uncertainty between EGC
and NEGC firms: linear regression and a propensity score matched sample. These approaches
rely on different assumptions and, thus, finding consistent results from both approaches increases
confidence in our inferences. The primary advantage of the linear regression approach is the
ability to estimate the mean difference in information uncertainty between the two groups of
firms while including a set of variables as controls for potential confounding effects. In addition,
the approach permits us to include all available firms, which increases the efficiency of our
estimation. The primary disadvantage of the linear regression approach is that it assumes
information uncertainty is a linear function of the explanatory variables. The primary advantage
of the propensity score matched sample approach is that it does not require specifying the
functional form of the relation between information uncertainty and the explanatory variables.
The disadvantage of this approach is the reduction in power arising from using a one-to-one
matched sample, i.e., limiting the number of NEGC firms to match the number of EGC firms.
In addition to testing for differences in information uncertainty between EGC and NEGC
firms, we also test whether any differences are related to the specific provisions of the JOBS Act
that EGC firms apply, e.g., confidentially filing draft IPO registration statements. Finding that
specific provisions of the JOBS Act are associated with differences in information uncertainty
between the two groups of firms increases our confidence that the differences are attributable to
the JOBS Act.
8
We do not compare information uncertainty in firms with EGC status to that of firms ineligible for EGC status, i.e.,
firms with more than $1 billion in pre-IPO revenue, for two reasons. First, IPO firms with more than $1 billion in
pre-IPO revenue likely are fundamentally different from the general population of IPO firms. Second, the revenue
threshold used to define EGC status is sufficiently high that it encompasses all but a few IPOs. See Section 4.
9
3.2 Do Firms with EGC Status Have Greater Information Uncertainty?
3.2.1 Linear Regression Approach
Using the linear regression approach, we test our predictions by estimating equation (1),
where the i subscript refers to firm.
InfoUncertaintyi = φ EGCi + δ1 Assetsi + δ2 Revenuei + δ3 BMi + δ4 ROAi
(1)
+ δ5 Agei + δ6 Techi + δ7 PctRetainedi + δ8 Big4i + γ Fixed Effectsi + ε
InfoUncertainty is either IPO underpricing or post-IPO equity return volatility. The focus of our
tests is the coefficient on EGC, which is an indicator variable that equals one for firms with EGC
status, and zero otherwise. The coefficient on EGC represents the mean difference in
information uncertainty between EGC and NEGC firms after controlling for the remaining
explanatory variables. Based on our prediction that the JOBS Act increased information
uncertainty for firms with EGC status, we predict φ > 0.
We measure IPO underpricing over three intervals. Following prior research, we
measure underpricing as the market-adjusted return from the IPO offer price to the closing price
on the day of the IPO, i.e., day t = 0, Underpricing(0). Because information uncertainty delays
the speed at which information is impounded in stock prices (e.g., Zhang, 2006), greater
information uncertainty could lead to delayed pricing. Accordingly, we also measure
underpricing using the closing price on the first trading day after the IPO, i.e., day t = 1,
Underpricing(0,1), and the closing price thirty trading days after the IPO, i.e., day t = 30,
Underpricing(0,30).9 We market-adjust our measures of underpricing as a means to control for
any systematic differences in market conditions between the two groups of firms.10
9
Using a thirty-day window also ensures that the closing price we use to calculate underpricing is not affected by
short-term price supports provided by the underwriter (e.g., Lowry, Officer, and Schwert, 2010).
10
Using raw returns or log-returns does not affect our inferences.
10
We use three measures of post-IPO volatility. The first, Volatility, is the standard
deviation of daily equity returns over the thirty-day window beginning the day after the IPO, i.e.,
t = 1 to 30. We exclude the IPO-day return, i.e., t = 0, to mitigate the effects of the large firstday returns on the volatility measure. The second and third measures separate Volatility into its
idiosyncratic and systematic components, IdiosyncVol and Beta. IdiosyncVol (Beta) is the
standard deviation of residuals (slope coefficient) from a firm-specific market model estimated
over the thirty-day window. Because we expect the reduction in information at a firm’s IPO
primarily to be idiosyncratic, separating volatility into idiosyncratic and systematic components
should increase the power of our tests; if information uncertainty is idiosyncratic, EGC should
have no association with Beta.11
Equation (1) includes control variables identified in prior research as determinants of IPO
underpricing and post-IPO volatility. In particular, we include as control variables firm size,
revenue, the equity book-to-market ratio, profitability, age, industry, ownership retention, auditor
quality, venture backing, exchange listing, and underwriter reputation (Loughran and Ritter,
2004; Purnanandam and Swaminathan, 2004; Schrand and Verrecchia, 2005; Leone, Rock, and
Willenborg, 2007; Lowry, Officer, and Schwert, 2010).
Assets is the natural logarithm of one plus total assets, Revenue is the natural logarithm of
one plus revenue, BM is book value of equity divided by market value of equity based on the
IPO offer price, and ROA is net income scaled by total assets. All accounting variables relate to
the fiscal year preceding the IPO. Age is the natural logarithm of one plus the number of years
from founding, or incorporation if the founding date is unavailable, to the date of the IPO. Tech
11
Lambert, Leuz, and Verrecchia (2007) shows that a reduction in disclosure of systematic (idiosyncratic)
information increases investors’ assessments of systematic (idiosyncratic) risk. In addition, using idiosyncratic
volatility ensures our results are not attributable to uncertainty about market-wide conditions or fluctuations in the
market return (Edelen and Kadlec, 2005; Pastor and Veronesi, 2005). 11
is an indicator variable that equals one if the firm is in a high technology industry based on the
Loughran and Ritter (2004) industry classification, and zero otherwise. PctRetained is the
percent of post-IPO shares outstanding retained by the pre-IPO shareholders. Big4 is an
indicator variable that equals one if the IPO firm’s auditor is Deloitte, EY, KPMG, or PwC, and
zero otherwise. Fixed Effects denotes fixed effects based on the firm’s Fama and French (1997)
twelve-industry group, the firm’s exchange listing, i.e., NYSE, NASDAQ, or AMEX, whether
the firm received venture backing, and the identity of the firm’s lead underwriter.12
Although the control variables are not central to our research question and we have no
predictions for the signs of their coefficients, prior research provides a basis for expected signs
on several of them. In particular, Ibbotson, Sindelar, and Ritter (1988), Beatty (1989), Loughran
and Ritter (2004), and Leone, Rock, and Willenborg (2007) find underpricing is negatively
related to firm size, the equity book-to-market ratio, and firm age, and positively related to return
on assets and whether the firm is in a high technology industry. Titman and Trueman (1986),
Grinblatt and Hwang (1989), Datar, Feltham, and Hughes (1991), and Schrand and Verrecchia
(2005) show that percentage of ownership retained by insiders and auditor quality are negatively
related to underpricing.
We estimate equation (1) using two samples, both of which comprise IPOs that meet the
JOBS Act revenue threshold. The first is the full sample of EGC and NEGC firms from January
2004 through December 2013. However, because this sample period includes the recent
financial crisis, it is possible that any difference in information uncertainty between EGC and
NEGC firms could be attributable to the financial crisis. Thus, we also estimate equation (1)
using the subsample of EGC and NEGC firms with IPOs after the financial crisis. We refer to
12
We include both Tech and industry fixed effects because Tech is not a proper subset of the twelve Fama and
French (1997) industry groups. When clustering our standard errors by industry, we use two-digit SIC codes to
ensure a sufficiently large number of clusters.
12
this as the post-Crisis sample. We base reported t-statistics on standard errors clustered by
industry and IPO date.
3.2.2 Propensity Score Matched Sample Approach
In the propensity score matched sample tests, we match each firm with EGC status to a
firm without EGC status with similar characteristics. To construct the matched sample, we first
estimate a propensity score for each firm as a function of the control variables in equation (1).
Specifically, we estimate EGC as a function of the control variables; the predicted values from
this regression are the propensity scores. We then match each EGC firm to the corresponding
NEGC firm that minimizes the squared difference in propensity score between the two firms.
This procedure results in a one-to-one match between EGC firms, i.e., the Treatment Sample,
and NEGC firms, i.e., the Propensity Score Matched Sample.13
We assess the validity of our matching procedure by testing for covariate balance
between the Treatment and Propensity Score Matched Samples. Specifically, we test for crosssample differences in the means and medians of the control variables used to calculate the
propensity score. A successful match is one in which these differences are insignificant. For a
successful match, we interpret any difference in information uncertainty between the two
samples as being attributable to the JOBS Act because the only known difference between the
two samples is the timing of their IPOs relative to the effective date of the Act (Rosenbaum,
2002; Armstrong, Jagolinzer, and Larcker, 2010).
3.3 Provisions of the JOBS Act
We next test whether any differences in information uncertainty between EGC and
NEGC firms are related to the specific provisions of the Act that the firms with EGC status
13
We select firms without EGC status using the full sample. Untabulated findings reveal that our inferences are
unaffected if we restrict these firms to those with IPOs after the financial crisis.
13
apply. Finding that specific provisions of the Act are associated with differences in IPO
underpricing and post-IPO volatility between the two groups of firms increases our confidence
that the differences are attributable to the JOBS Act.
We base the analysis in this section on information provided to us by EY regarding the
specific JOBS Act provisions each EGC firm applies. The information provided by EY indicates
whether each firm with EGC status (1) filed its IPO draft registration statements confidentially
with the SEC, (2) provided summary compensation information for fewer than five top
executives, (3) omitted the compensation discussion and analysis, (4) provided less than three
years of audited financial statements, (5) elected to delay application of new or revised
accounting standards, and (6) elected to delay the internal control audit required by Section
404(b) of SOX. For each of these six provisions we construct an indicator variable that equals
one if the firm applies the provision, and zero otherwise.
We use two approaches to test for a relation between these provisions and the differences
in IPO underpricing and post-IPO volatility between EGC and NEGC firms. First, we create an
index, JOBSDiscIndex, that equals the sum of the six provision indicator variables, and include
this index as an additional explanatory variable in equation (1). Because each indicator variable
equals one if the provision reduces disclosure, we expect that the larger is the index the greater is
information uncertainty. Thus, we predict a positive relation between JOBSDiscIndex and IPO
underpricing and post-IPO volatility. Second, rather than use an index, we include the indicator
variables in equation (1). We predict a positive relation between each indicator variable and IPO
underpricing and post-IPO volatility. However, because we have no basis to predict which
provisions have larger or smaller effects on information uncertainty, we make no predictions
about the relative magnitudes of the relations.
14
An advantage of using the index is that our inferences depend only on the number of
JOBS Act provisions an EGC firm applies, rather than which specific provision it applies. This
is particularly relevant because few observations identify the effects of some provisions. For
example, 96% of EGC firms that file their draft IPO registration statements confidentially also
omit the compensation discussion and analysis. The advantage of using the separate indicators is
that doing so allows us to estimate the effects on information uncertainty associated with specific
provisions.
4. Sample and Data
We construct our sample by identifying all IPOs in the US between January 1, 2004 and
December 31, 2013 that appear in the Securities Data Corporation (SDC) database, list common
stock on the NYSE, NASDAQ, or AMEX, file registrations statements on form S-1 with the
SEC, and have the stock price and accounting information necessary for our tests. From SDC we
obtain the IPO date, the IPO issue price, the number of shares offered and outstanding after the
IPO, whether the firm was venture backed, the lead underwriter, and the firm’s auditor. We
obtain share price, equity returns, the market return, industry, and exchange listing from CRSP,
revenue, net income, total assets, and book value of equity from Compustat, and firm age from
the Field-Ritter dataset of company founding dates (Field and Karpoff, 2002; Loughran and
Ritter, 2004).14 We exclude IPOs that are leveraged buyouts, closed-end funds, open-end funds,
trusts, and special purpose vehicles (i.e., SIC codes 6091, 6371, 6722, 6726, 6732, 6733, and
6799).
There are 207 firms with IPOs after the JOBS Act, i.e., after April 5, 2012, that meet
these data requirements. To construct our sample of EGC firms, from among these 207 firms we
exclude all 19 firms that had IPOs in April and May 2012 that had filed their prospectuses before
14
The Field-Ritter dataset is available on Jay Ritter’s Web page (http://bear.cba.ufl.edu/ritter/FoundingDates.htm).
15
the JOBS Act took effect, 28 firms that did not qualify for EGC status, i.e., had gross revenues in
excess of $1 billion in the year prior to the IPO, and 2 firms that qualified for EGC status but did
not elect to apply any provision of the JOBS Act.15,16 The resulting sample of EGC firms
comprises 158 firms that meet the JOBS Act revenue threshold and have an IPO between June
2012 and December 2013.
Our sample of NEGC firms comprises 775 firms that meet the JOBS Act revenue
threshold and have IPOs between January 2004 and March 2012, i.e., before the JOBS Act
became effective. To mitigate concerns that differences in information uncertainty between
EGC and NEGC firms are attributable to the financial crisis, we also employ a sample of postCrisis NEGC firms. The post-Crisis sample comprises the 218 NEGC firms that have IPOs
beginning July 2009.17 Hence, the full sample comprises 933 firms (158 EGC firms + 775
NEGC firms), and the post-Crisis sample comprises 376 firms (158 EGC firms + 218 NEGC
firms).
Table 1 presents descriptive statistics for the full sample of 933 firms. Table 1 reveals
that mean first-day underpricing, Underpricing(0), is 14.41%. This finding is consistent with
prior research in that it indicates that IPOs are, on average, underpriced. Table 1 also reveals that
mean underpricing over the longer horizons, Underpricing(0,1) and Underpricing(0,30), are
15.25% and 16.14%. These statistics indicate that underpricing persists for several weeks
beyond the first day, which is consistent with findings in Lowry, Officer, and Schwert (2010).
Mean Volatility, IdiosyncVol, and Beta are 3.02, 2.88, and 0.56. These statistics indicate that
most of the post-IPO return volatility is idiosyncratic, and that returns shortly after the IPO
15
Three of the 19 firms that had IPOs in April and May 2012 exceed the $1 billion revenue threshold and are
ineligible for EGC status, and the remaining 16 elected not to apply many of the provisions of the Act.
16
Inferences are unaffected if any or all of the firms in these three groups are classified as NEGC firms.
17
We begin the post-Crisis sample in July 2009 because NBER business cycle dates indicate that a recession starts
in the fourth quarter of 2007 and continues until June 2009.
16
exhibit little co-movement with the market, which helps rule out changes in market conditions as
an alternative explanation for our findings.
Table 1 also indicates that, on average, IPO firms report large losses in the year preceding
the IPO (mean ROA = –0.22), approximately 19% of IPO firms are in high technology industries
(mean Tech = 0.19), pre-IPO investors retain approximately 65% of shares outstanding (mean
PctRetained = 0.65), and 79% of firms use a Big 4 auditor to audit the financial statements
included in the IPO registration statement (mean Big4 = 0.79).
Table 2 presents descriptive statistics separately for EGC and NEGC firms. Panel A
(Panel B) presents means and medians of variables used in our analysis for the full (post-Crisis)
sample. Panel A indicates that for NEGC and EGC firms, means (medians) of Underpricing(0)
are 13.14% and 20.64% (7.51% and 14.28%), means (medians) of Underpricing(0,1) are 13.78%
and 22.47% (% and 14.29%), and means (medians) of Underpricing(0,30) are 13.90% and
27.14% (% and 19.74%). For NEGC and EGC firms, means (medians) of Volatility are
2.97% and 3.28% (2.63% and 3.06%), means (medians) of IdiosyncVol are 2.83% and 3.17%
(2.52% and 2.89%), and means (medians) of Beta are 0.56 and 0.56 (0.51 and 0.46). All of the
differences in means and medians between NEGC and EGC firms are significant, except for
those relating to Beta.18 For example, the difference in mean (median) for Underpricing(0) is
7.50% = 20.64%  13.14% (6.77% = 14.28%  7.51%). In addition, the means and medians for
the underpricing measures for EGC firms are approximately 1.5 to 2 times those of NEGC firms.
Table 2, panel A, also indicates that EGC firms generally are smaller, less profitable,
younger, and less likely to be in high technology industries. For example, the p-values for
differences in means between firms EGC and NEGC firms for Revenue, ROA, Age, and Tech are
18
Throughout, we use the term significant to denote a five percent significance level under a two-sided alternative.
17
0.05, < 0.001, 0.003, and 0.002. The mean differences in assets, Assets, the equity book-tomarket ratio, BM, the percent of shares retained by pre-IPO investors, PctRetained, and audit
quality, Big4, are not significantly different between the two groups of firms.
Panel B presents statistics analogous to those in panel A, but based on IPOs after the
financial crisis, i.e., the post-Crisis sample. Inferences generally are similar to those based on the
full sample statistics. As in panel A, panel B indicates larger IPO underpricing and higher postIPO volatility for EGC firms. For example, the mean (median) difference in Underpricing(0)
between the two groups of firms is 13.21% (6.70%) vs. 20.64% (14.28%). As in panel A, the
statistics in panel B also indicate that EGC firms generally are smaller, less profitable, and
younger.
5. Results
5.1 Linear Regression
Table 3 presents results from estimating equation (1) when the dependent variable is one
of the three underpricing measures, Underpricing(0), Underpricing(0,1), and
Underpricing(0,30). Panel A (Panel B) presents results for the full (post-Crisis) sample. Table 3
reveals that EGC firms have significantly larger underpricing than NEGC firms. In panel A, the
coefficients on EGC when Underpricing(0), Underpricing(0,1), and Underpricing(0,30) are the
dependent variables are 6.52, 6.63, and 13.98 (t-statistics = 6.05, 6.48, and 8.75). Regarding the
control variables, panel A indicates that IPO underpricing generally is larger for firms with
higher revenue, smaller for firms with higher equity book-to-market ratios, and larger for firms
with higher return on assets. Inferences based on the findings in panel B are similar.
Specifically, the coefficients on EGC are 6.63, 7.14, and 14.14 (t-statistics = 5.54, 7.88, and
5.22). These findings are consistent with the JOBS Act increasing IPO underpricing.
18
Table 4 presents results from estimating equation (1) when the dependent variable is one
of the three post-IPO volatility measures, Volatility, IdiosyncVol, and Beta. Panel A (Panel B)
presents results for the full (post-Crisis) sample. Table 4 reveals that EGC firms have
significantly higher conditional mean Volatility and IdiosyncVol than NEGC firms and have
insignificantly different Beta. In panel A, the coefficients on EGC are 0.27, 0.30, and –0.06 (tstatistics = 2.45, 2.90, and –1.13). These findings are consistent with the JOBS Act increasing
post-IPO volatility and with most of the increase in volatility being idiosyncratic. Regarding the
control variables, Table 4 indicates that smaller, growth, and younger firms have higher
volatility. Inferences based on the findings in panel B are similar. Specifically, the coefficients
on EGC are 0.27, 0.32, and –0.08 (t-statistics = 3.00, 3.63, and –1.53). These findings are
consistent with the JOBS Act increasing post-IPO return volatility.
5.2 Propensity Score Matched Sample
Table 5 presents results from the propensity score matched sample tests. Panel A
presents means and medians for the variables we use to estimate propensity scores for the
Treatment Sample and Propensity Score Matched Sample. We find no significant differences
between the two samples in means and medians for the eight control variables, with the single
exception of the difference in medians for PctRetained (p-value = 0.10). These findings are
consistent with the propensity score matching resulting in covariate balance and hence a
successful match. In addition, the absolute magnitudes of the differences in means and medians
are small. Finding that fifteen of the sixteen means and medians of the control variables do not
differ increases our confidence that any differences in our measures of information uncertainty
between the Treatment and Propensity Score Matched Samples are attributable to the JOBS Act.
19
Panel B presents the differences between the two samples in means and medians of our
measures of information uncertainty. Panel B reveals significantly larger mean and median IPO
underpricing and significantly higher mean and median post-IPO volatility for EGC firms than
their matched-sample counterparts. For Underpricing(0), Underpricing(0,1), and
Underpricing(0,30), the two-tailed p-values for the differences in means and medians range from
< 0.001 to 0.01, and for Volatility and IdiosyncVol, they range from 0.02 to 0.06. There is no
significant difference in mean or median Beta (p-values = 0.88 and 0.46). The inferences based
on the propensity score matched sample test results are consistent with those based on the linear
regression results, and are consistent with the JOBS Act increasing underpricing and return
volatility in the IPO market.
5.3 Provisions of the JOBS Act
Table 6 presents descriptive statistics for indicator variables relating to the provisions of
the JOBS Act for the 158 EGC firms. Confidential equals one if the firm filed confidentially its
draft registration statement with the SEC. ReduceComp equals one if the firm presented
compensation information for fewer than five top executives. OmitCDA equals one if the firm
omitted the compensation discussion and analysis. ReduceAcct equals one if the firm presented
only two years of audited financial statements. DelayGAAP equals one if the firm elected to
delay application of new or revised accounting standards. DelaySOX equals one if a firm
delayed the internal controls audit required by Section 404(b) of SOX. Each indicator variable
equals zero otherwise.
Panel A presents descriptive statistics relating to the number of firms electing each
provision. Panel A reveals that 73% of the firms file their draft registration statements
confidentially with the SEC (mean Confidential = 0.73), which suggests that firms view public
20
disclosure about a potential upcoming IPO as costly. Panel A also reveals that 84% elect to
present compensation information for fewer than five top executives (mean ReduceComp = 0.84)
and 87% omit the compensation discussion and analysis (mean OmitCDA = 0.87). These
statistics suggest the vast majority of firms perceive that the cost of these disclosures outweighs
the benefit. In contrast, only 43% of the firms present fewer than three years of audited financial
statements (mean ReduceAcct = 0.43) and only 18% elect to delay application of new or revised
accounting standards (mean DelayGAAP = 0.18). The fact that the vast majority of firms
provide less compensation information, but a minority of firms provide less accounting
information suggests that firms perceive the net benefits of disclosing accounting information as
being higher than the net benefits of disclosing compensation information. Finally, panel A
reveals that 100% of the firms elect to delay the internal controls audit required by SOX (mean
DelaySOX = 1.00). Finding that no firm voluntarily complies with Section 404(b) of SOX is
consistent the firms viewing the costs of compliance as exceeding the benefits (Zhang, 2007;
Gao, Wu, and Zimmerman, 2009).
Panel B presents distributional statistics for the disclosure index, JOBSDiscIndex, which
we construct as the sum of the six provision variables in panel A. Because the index includes
DelaySOX, JOBSDiscIndex ranges from 1 to 6. Panel B reveals that, on average, firms apply
approximately four of the six provisions (mean and median JOBSDiscIndex = 4.05 and 4.00).
More than 25% of the firms apply five provisions (75th percentile of JOBSDiscIndex = 5.00).
Panel C reports that the correlations between the provision indicator variables are significant at
the 10% level or less, except for those relating to DelayGAAP. Not surprisingly, the two
compensation provisions, ReduceComp and OmitCDA, are highly correlated (corr. = 0.46).
OmitCDA and Confidential also are highly correlated (corr. = 0.45).
21
Table 7 presents the results from estimating a version of equation (1) that includes
JOBSDiscIndex as an additional explanatory variable. Panel A (Panel B) presents results for the
IPO underpricing (post-IPO volatility) specifications.19 The key finding in panel A is that the
JOBSDiscIndex coefficients are significantly positive in all three underpricing specifications (tstatistics = 3.11, 5.39, and 4.56), and the EGC coefficients are either insignificantly different
from zero or significantly negative (t-statistics = 0.22, 1.03, and 1.87). The key finding in
panel B is that the JOBSDiscIndex coefficients are significantly positive in the Volatility and
IdiosyncVol specifications (t-statistics = 4.56 and 5.62), and the EGC coefficients are
insignificantly different from zero (t-statistics = –1.23 and –1.68). Taken together, the results in
Table 7 suggest that variation in the extent to which IPO firms apply provisions of the JOBS Act
explains the significantly larger underpricing and higher post-IPO volatility for EGC firms;
simply having an IPO after the JOBS Act does not.
Table 8 presents the findings from estimating a version of equation (1) that includes five
provision indicator variables, Confidential, ReduceComp, OmitCDA, ReduceAcct, and
DelayGAAP, as additional explanatory variables rather than JOBSDiscIndex.20 Panel A (Panel
B) presents results for the IPO underpricing (post-IPO volatility) specifications. Panel A reveals
that the coefficients on Confidential and ReduceComp are significantly positive for all three
underpricing specifications. In particular, the Confidential coefficients (t-statistics) are 6.57,
8.73, and 12.71 (3.96, 4.44, and 3.05), and the ReduceComp coefficients (t-statistics) are 5.93,
8.79, and 13.51 (3.79, 5.97, and 6.97). Regarding the remaining provisions, panel A reveals that
eight of the nine coefficients on OmitCDA, ReduceAcct, and DelayGAAP are insignificantly
19
For parsimony we tabulate results for the full sample. Untabulated findings reveal the same inferences based on
the post-Crisis sample. Because DelaySOX equals one for all EGC firms, DelaySOX does not contribute to crosssectional variation in JOBSDiscIndex.
20
We do not include DelaySOX because all firms apply this provision. For the sake of parsimony, we do not
tabulate the control variable coefficients.
22
different from zero, and one is significantly negative. These findings suggest that IPO firms that
confidentially file their draft registration statements or present compensation information for
fewer than five top executives have first-day underpricing that is incrementally higher by
approximately 6%.
Panel B reveals that the Confidential, ReduceComp, OmitCDA, and ReduceAcct
coefficients are significantly positive for the IdiosyncVol specification (t-statistics = 1.73, 1.70,
2.48, and 2.23), and positive for the Volatility specification, significantly so for two of the four,
Confidential and ReduceAcct (t-statistics = 2.11, 1.31, 1.63, and 2.50). The coefficients range
from 0.20 to 0.31, which is consistent with firms that file confidentially their draft IPO
registration statements or provide reduced compensation or accounting disclosure having postIPO total and idiosyncratic volatility that is incrementally higher by approximately 0.25 per
provision.
6. Summary and Concluding Remarks
The question this study addresses is whether the JOBS Act increases information
uncertainty in firms with IPOs. The JOBS Act creates a new category of issuer, the Emerging
Growth Company (EGC), and eases regulations for EGCs to encourage initial public offerings of
their shares. Specifically, the Act allows EGC firms to reduce the scope of mandatory disclosure
of financial statement and executive compensation information, to file draft IPO registration
statements confidentially with the SEC, to delay application of new or revised accounting
standards, and to delay compliance with the Sarbanes-Oxley Act’s requirement for auditor
attestation on internal controls. We find evidence consistent with the easing of these regulations
increasing information uncertainty in the IPO market.
We address this research question by testing for differences in IPO underpricing and
post-IPO equity return volatility between IPO firms with EGC status and IPO firms that would
23
have qualified for EGC status had their IPO been after the JOBS Act. We also test whether
variation in the extent to which IPO firms apply provisions of the JOBS Act explains any
differences in underpricing and post-IPO volatility.
Results from these tests are consistent with our prediction that IPO firms with EGC status
have greater information uncertainty at the time of their IPOs. In particular, findings from the
linear regression and propensity score matched sample approaches reveal that firms with EGC
status have larger IPO underpricing and higher total and idiosyncratic post-IPO volatility.
Additionally, we find that cross-sectional variation in the number of JOBS Act provisions EGC
firms apply, as well as specific provisions that they apply, explains the differences in
underpricing and volatility; simply having an IPO after the JOBS Act does not. These findings
increase our confidence that the differences are attributable to the JOBS Act.
Taken together, the findings indicate that the JOBS Act increased information uncertainty
in IPO firms. These findings provide support for critics’ concerns that although the JOBS Act’s
eased disclosure requirements may decrease the compliance costs associated with filing IPO
registration statements, the Act increases the costs of an IPO by increasing the risk premium
investors demand to compensate them for the resulting increase in information uncertainty.
24
References
Armstrong, C., Jagolinzer, A., Larcker, D., 2010. Chief Executive Officer Equity Incentives and
Accounting Irregularities. Journal of Accounting Research 48: 225-271.
Arnold, T., Fishe, R., North, D., 2010. The Effects of Ambiguous Information on Initial and
Subsequent IPO Returns. Financial Management 39: 1497-1519.
Beatty, R., 1989. Auditor Reputation and the Pricing of Initial Public Offerings. The Accounting
Review 64: 693-709.
Beatty, R., Ritter, J., 1986. Investment Banking, Reputation, and the Underpricing of Initial
Public Offerings. Journal of Financial Economics 15: 213-232.
Boulton, T., Smart, S., Zutter, C., 2011. Earnings Quality and International IPO Underpricing.
The Accounting Review 86: 483-505.
Datar, S., Feltham, G., Hughes, J., 1991. The Role of Audits and Audit Quality in Valuing New
Issues. Journal of Accounting and Economics 14: 2-49.
Edelen, R., Kadlec, G., 2005. Comparable-firm Returns, Issuer Surplus, and the Pricing and
Withdrawal of IPOs. Journal of Financial Economics 77, 347–373.
Ernst & Young LLP, 2013. The JOBS Act: One-year Anniversary, April.
Ernst & Young LLP, 2013. The JOBS Act: 18 Months Later, November.
Fama, E., French, K., 1997. Industry Costs of Equity. Journal of Financial Economics 43: 153193.
Field, L., Karpoff, J., 2002. Takeover Defenses in IPO Firms. Journal of Finance 57:
1857‒1889.
Gao, F., Wu, J., Zimmerman, J., 2009. Unintended Consequences of Granting Small Firms
Exemptions from Securities Regulation: Evidence from the Sarbanes-Oxley Act. Journal of
Accounting Research 47, 459-506.
Grinblatt, M., Hwang, C., 1989. Signaling and the Pricing of New Issues. Journal of Finance 44:
393-420.
Ibbotson, R., Sindelar J., and Ritter, J., 1988. Initial Public Offerings. Journal of Applied
Corporate Finance 1: 37-45.
Lambert, R., Leuz, C., Verrecchia, R., 2007. Accounting Information, Disclosure, and the Cost
of Capital. Journal of Accounting Research 45: 385-420.
25
Latham and Watkins, 2013. The JOBS Act After One Year: A Review of the New IPO
Playbook.
Latham and Watkins, 2014. The JOBS Act, Two Years Later: An Updated Look at the IPO
Landscape.
Leone, A.J., Rock, S., Willenborg, M., 2007. Disclosure of Intended Use of Proceeds and
Underpricing in Initial Public Offerings. Journal of Accounting Research 45: 111-153.
Loughran, T., Ritter, J., 2004. Why Has IPO Underpricing Changed over Time? Financial
Management 33: 5-37.
Lowry, M., Officer, M., Schwert, G., 2010. The Variability of IPO Initial Returns. Journal of
Finance 65: 425-465.
Ljungqvist, A., 2007. IPO Underpricing. in Eckbo, B., (Ed.), Handbook of Corporate Finance.
Elsevier, Amsterdam.
Pastor, L., Veronesi, P., 2005. Rational IPO waves. Journal of Finance 58, 1749–1789.
Purnanandam, A. Swaminathan, B., 2004. Are IPOs Really Underpriced? Review of Financial
Studies 17: 811-848.
Ritter, J., Welch, I., 2002. A Review of IPO Activity, Pricing and Allocations. Journal of
Finance 57: 1795-1828.
Rock, K., 1986. Why New Issues are Underpriced? Journal of Financial Economics 15: 187212.
Rosenbaum, P., 2002. Observational Studies, second ed. Springer Series in Statistics, Berlin.
Schrand, C., Verrecchia, R., 2005. Information Disclosure and Adverse Selection Explanations
for IPO Underpricing. Working paper, University of Pennsylvania.
Titman, S., Trueman, B., 1986. Information Quality and the Valuation of New Issues. Journal of
Accounting and Economics 8: 159-172.
Willenborg, M., 1999. Empirical Analysis of the Economic Demand for Auditing in the Initial
Public Offerings Market. Journal of Accounting Research 37, 225-238.
Zhang, F. 2006. Information Uncertainty and Stock Returns. Journal of Finance 61, 105-137.
Zhang, I. 2007. Economic Consequences of the Sarbanes-Oxley Act of 2002. Journal of
Accounting and Economics 44, 74-115.
26
Table 1. Descriptive Statistics
Variable
Underpricing(0)
Underpricing(0,1)
Underpricing(0,30)
Volatility
IdiosyncVol
Beta
Assets
Revenue
BM
ROA
Age
Tech
PctRetained
Big4
Mean
14.41
15.25
16.14
3.02
2.88
0.56
4.68
3.99
0.18
–0.22
2.49
0.19
0.65
0.79
Std
21.83
23.61
28.65
1.56
1.52
1.00
1.55
1.84
0.30
0.69
0.85
0.39
0.22
0.41
25th
–0.12
–0.24
–2.32
1.87
1.77
0.01
3.61
3.19
0.02
–0.30
1.95
0.00
0.61
1.00
Median
7.97
9.10
11.14
2.69
2.57
0.51
4.51
4.35
0.10
0.00
2.30
0.00
0.72
1.00
75th
23.47
24.69
28.82
3.92
3.75
1.10
5.79
5.35
0.24
0.06
3.00
0.00
0.79
1.00
This table presents descriptive statistics for variables used in our tests. Underpricing(0) is the
return on the first trading day, i.e., closing price on the day of the IPO minus the IPO offer price
divided by the offer price, Underpricing(0,1) is the return from the offer price to the closing
price on the second trading day, and Underpricing(0,30) is the return from the offer price to the
closing price thirty trading days later. Underpricing(0), Underpricing(0,1), and
Underpricing(0,30) are in excess of the market return over the respective window and expressed
in percent. Volatility is the standard deviation of daily returns over the thirty-day window
beginning the day after the IPO, i.e., t = 1 to 30, and IdiosyncVol (Beta) is the standard deviation
of residuals (slope coefficient) from a firm-specific market model estimated over the same
window. Assets (Revenue) is the natural logarithm of one plus total assets (revenue) in the year
precedingthe IPO. BM is equity book value in the year preceding the IPO scaled by equity
market value based on the offer price. ROA is net income scaled by assets, both for the year
preceding the IPO. Age is the natural logarithm of one plus the number of years from founding
or incorporation, if the founding date is unavailable, to IPO. Tech is an indicator variable that
equals one if the IPO firm is a technology company based on the Loughran and Ritter (2004)
classification. PctRetained is the percent of post-IPO shares outstanding retained by the pre-IPO
shareholders. Big4 is an indicator variable that equals one if the firm’s auditor is Deloitte, EY,
KPMG, or PwC. The sample comprises 933 IPO firms from 2004 to 2013.
27
Table 2. Univariate Analysis of Capital Market Outcomes
Panel A. Means and Medians: Full Sample
Variable
Underpricing(0)
Underpricing(0,1)
Underpricing(0,30)
Volatility
IdiosyncVol
Beta
Assets
Revenue
BM
ROA
Age
Tech
PctRetained
Big4
NEGC Firms
(N = 775)
mean median
13.14
7.51
13.78
7.99
13.90
9.51
2.97
2.63
2.83
2.52
0.56
0.51
4.71
4.59
4.04
4.38
0.18
0.10
–0.18
0.01
2.53
2.40
0.21
0.00
0.66
0.72
0.79
1.00
EGC Firms
(N = 158)
mean median
20.64
14.28
22.47
14.29
27.14
19.74
3.28
3.06
3.17
2.89
0.56
0.46
4.52
4.38
3.73
4.23
0.21
0.08
–0.43
–0.08
2.31
2.20
0.10
0.00
0.64
0.74
0.79
1.00
p-value: test
of difference
in means
<0.001
<0.001
<0.001
0.02
0.01
0.98
0.14
0.05
0.28
<0.001
0.003
0.002
0.29
0.98
p-value: test
of difference
in medians
0.004
0.002
<0.001
0.02
0.01
0.82
0.21
0.06
0.11
<0.001
0.01
NA
0.29
NA
p-value: test
of difference
in means
0.004
0.001
<0.001
0.05
0.02
0.87
0.04
0.06
0.45
0.02
0.15
0.64
0.74
0.12
p-value: test
of difference
in medians
0.02
0.01
<0.001
0.02
0.01
0.56
0.03
0.05
0.05
0.002
0.09
NA
0.47
NA
Panel B. Means and Medians: Post-Crisis Sample
Variable
Underpricing(0)
Underpricing(0,1)
Underpricing(0,30)
Volatility
IdiosyncVol
Beta
Assets
Revenue
BM
ROA
Age
Tech
PctRetained
Big4
NEGC Firms
(N = 218)
mean median
13.21
6.70
13.51
7.40
13.37
9.40
2.94
2.59
2.77
2.38
0.57
0.55
4.84
4.83
4.10
4.42
0.24
0.12
–0.22
0.00
2.44
2.48
0.09
0.00
0.63
0.72
0.72
1.00
EGC Firms
(N = 158)
mean median
20.64
14.28
22.47
14.29
27.14
19.74
3.28
3.06
3.17
2.89
0.56
0.46
4.52
4.38
3.73
4.23
0.21
0.08
–0.43
–0.07
2.31
2.20
0.10
0.00
0.64
0.74
0.79
1.00
This table reports means and medians, for NEGC and EGC firms, of variables we use in our
tests. Panel A presents statistics for the full sample, and panel B presents statistics for the
28
sample of firms with IPOs after the financial crisis, i.e., beginning July 2009. p-values are for
two-tailed tests of differences in mean and medians. See Table 1 for variable definitions.
29
Table 3. Underpricing
Panel A. Full Sample
Variable
EGC
Assets
Revenue
BM
ROA
Age
Tech
PctRetained
Big4
Fixed
Effects
R2 (%) / N
Underpricing(0)
coef
t-stat
6.52
6.05
–1.19
–1.46
1.98
3.64
–5.02
–2.18
3.37
4.66
–1.71
–2.19
2.24
1.06
–5.04
–1.45
–0.80
–0.67
Industry, Exchange,
Venture, Underwriter
20.07 / 933
Underpricing(0,1)
coef
t-stat
6.63
6.48
–2.01
–1.98
2.22
4.79
–3.70
–2.10
4.38
7.54
–1.75
–2.56
–0.56
–0.20
–2.30
–0.73
–0.30
–0.25
Industry, Exchange,
Venture, Underwriter
18.61 / 933
Underpricing(0,30)
coef
t-stat
13.98
8.75
–2.49
–2.17
1.93
2.12
–2.98
–1.48
5.98
5.54
–1.27
–0.96
1.00
0.32
–3.12
–0.88
2.85
1.73
Industry, Exchange,
Venture, Underwriter
17.00 / 933
Underpricing(0,1)
coef
t-stat
7.14
7.88
–0.38
–0.48
1.46
3.82
–3.93
–2.20
4.33
3.96
–1.91
–2.20
–0.18
–0.02
–6.95
–0.88
2.93
1.29
Industry, Exchange,
Venture, Underwriter
30.05 / 376
Underpricing(0,30)
coef
t-stat
14.14
5.22
0.32
0.30
–1.37
–2.44
–4.29
–2.60
7.31
5.46
0.96
0.79
18.16
1.91
–4.77
–0.71
4.37
1.25
Industry, Exchange,
Venture, Underwriter
27.82 / 376
Panel B. Post-Crisis Sample
Variable
EGC
Assets
Revenue
BM
ROA
Age
Tech
PctRetained
Big4
Fixed
Effects
R2 (%) / N
Underpricing(0)
coef
t-stat
6.63
5.54
0.56
0.70
0.81
1.48
–5.17
–1.79
3.74
4.71
–1.42
–1.95
1.13
0.12
–7.15
–0.90
2.13
1.40
Industry, Exchange,
Venture, Underwriter
32.26 / 376
This table presents results from estimating a linear regression of IPO underpricing on EGC, an
indicator that equals one for EGC firms, and control variables. Panel A presents results for the
full sample, and panel B presents results for the sample of firms with IPOs after the financial
crisis, i.e., beginning July 2009. t-statistics are based on standard errors clustered by industry
and IPO date. See Table 1 for variable definitions.
30
Table 4. Return Volatility
Panel A. Full Sample
Variable
EGC
Assets
Revenue
BM
ROA
Age
Tech
PctRetained
Big4
Fixed
Effects
R2 (%) / N
Volatility
coef
t-stat
0.27
2.45
–0.21
–4.04
–0.02
–0.81
–0.56
–3.44
–0.07
–0.80
–0.14
–1.87
–0.07
–0.50
0.58
1.73
0.13
1.24
Industry, Exchange,
Venture, Underwriter
27.48 / 933
IdiosyncVol
coef
t-stat
0.30
2.90
–0.21
–3.98
–0.02
–0.71
–0.37
–2.93
–0.08
–0.89
–0.14
–1.90
0.01
0.04
0.58
1.66
0.16
1.16
Industry, Exchange,
Venture, Underwriter
26.94 / 933
Beta
coef
t-stat
–0.06
–1.13
–0.06
–1.93
0.06
1.97
–0.07
–0.62
–0.04
–0.36
–0.06
–1.96
0.08
0.82
0.03
0.22
–0.04
–0.50
Industry, Exchange,
Venture, Underwriter
6.07 / 933
Panel B. Post-Crisis Sample
Variable
EGC
Assets
Revenue
BM
ROA
Age
Tech
PctRetained
Big4
Fixed
Effects
R2 (%) / N
Volatility
coef
t-stat
0.27
3.00
–0.33
–6.02
–0.03
–1.03
–0.23
–1.71
0.05
0.60
–0.09
–1.62
–1.10
–9.03
0.72
1.38
0.31
2.88
Industry, Exchange,
Venture, Underwriter
35.13 / 376
IdiosyncVol
coef
t-stat
0.32
3.63
–0.30
–6.48
–0.02
–0.80
–0.20
–1.53
0.01
0.03
–0.10
–1.83
–0.98
–5.79
0.74
1.43
0.36
3.69
Industry, Exchange,
Venture, Underwriter
34.32 / 376
Beta
coef
t-stat
–0.08
–1.53
–0.13
–2.80
0.04
0.81
0.18
3.11
0.11
4.84
–0.04
–1.70
–0.67
–2.73
0.20
1.36
–0.05
–0.40
Industry, Exchange,
Venture, Underwriter
11.15 / 376
This table presents results from estimating a linear regression of post-IPO equity return volatility
on EGC, an indicator that equals one for EGC firms, and control variables. Panel A presents
results for the full sample, and panel B presents results for the sample of firms with IPOs after
the financial crisis, i.e., beginning July 2009. t-statistics are based on standard errors clustered
by industry and IPO date. See Table 1 for variable definitions.
31
Table 5. Propensity Score Matched Sample
Panel A. Differences in Control Variables
Variable
Assets
Revenue
BM
ROA
Age
Tech
PctRetained
Big4
Propensity Score
Matched Sample
(N = 158)
mean median
4.42
4.19
3.78
4.05
0.21
0.09
–0.39 –0.01
2.24
2.08
0.15
0.00
0.64
0.71
0.76
1.00
Treatment
Sample
(N = 158)
mean median
4.52
4.38
3.73
4.23
0.21
0.08
–0.43 –0.08
2.31
2.20
0.10
0.00
0.64
0.74
0.79
1.00
p-value: test
of difference
in means
0.60
0.84
0.93
0.74
0.44
0.18
0.96
0.54
p-value: test
of difference
in medians
0.53
0.93
0.52
0.31
0.11
NA
0.10
NA
p-value: test
of difference
in means
0.003
0.001
<0.001
0.03
0.02
0.88
p-value: test
of difference
in medians
0.01
0.01
<0.001
0.06
0.04
0.46
Panel B. Differences in Underpricing and Return Volatility
Variable
Underpricing(0)
Underpricing(0,1)
Underpricing(0,30)
Volatility
IdiosyncVol
Beta
Propensity Score
Matched Sample
(N = 158)
mean median
11.36
6.20
11.80
8.29
11.88 11.22
2.90
2.62
2.77
2.50
0.57
0.55
Treatment
Sample
(N = 158)
mean median
20.64 14.28
22.47 14.29
27.14 19.74
3.28
3.06
3.17
2.89
0.56
0.46
This table presents results from using propensity score matching to match EGC firms to NEGC
firms. One-to-one matched pairs are formed by estimating a propensity score as a function of
control variables, and minimizing the differences in propensity scores. Panel A presents
differences in control variables between EGC firms, i.e., Treatment Sample, and propensity score
matched sample of NEGC firms, i.e., Propensity Score Matched Sample. Panel B presents
differences in IPO underpricing and post-IPO return volatility between the Treatment and
Propensity Scored Matched Samples. p-values are for two-tailed tests of differences in mean and
medians. See Table 1 for variable definitions.
32
Table 6. Additional Analysis––JOBS Act Provisions
ReduceComp
OmitCDA
ReduceAcct
DelayGAAP
DelaySOX
Value
N (Variable = 1)
N (Variable = 0)
Mean
Confidential
Panel A. Descriptive Statistics
115
43
0.73
133
25
0.84
138
20
0.87
68
90
0.43
28
130
0.18
158
0
1.00
Panel B. Provision Index
Variable
Mean
4.05
JOBSDiscIndex
Std
1.17
Min
1.00
10th
2.00
25th
4.00
Median
4.00
75th
5.00
90th
5.00
Max
6.00
0.25
0.03
0.58
0.16
0.17
0.25
–0.09
–0.07
0.03
0.03
0.03
0.71
JOBSDiscIndex
DelayGAAP
0.46
0.17
–0.07
0.49
0.45
0.46
ReduceAcct
0.28
0.28
0.45
0.16
–0.09
0.59
OmitCDA
ReduceComp
Variable
Confidential
ReduceComp
OmitCDA
ReduceAcct
DelayGAAP
JOBSDiscIndex
Confidential
Panel C. Correlation Matrix
0.64
0.60
0.72
0.62
0.30
0.33
This table presents descriptive statistics regarding provisions of the JOBS Act that EGC firms
applied at their IPOs. Confidential equals one if the firm filed confidentially its draft registration
statement with the SEC. ReduceComp equals one if the firm presented compensation
information for fewer than five top executives. OmitCDA equals one if the firm omitted the
compensation discussion and analysis. ReduceAcct equals one if the firm presented only two
years of audited financial statements. DelayGAAP equals one if the firm elected to delay
application of new or revised accounting standards. DelaySOX equals one if a firm delayed the
internal controls audit required by Section 404(b) of SOX. Each indicator variable equals zero
otherwise. JOBSDiscIndex is the sum of the six provision variables. Panel A presents number of
firms electing each provision. Panel B presents descriptive statistics for JOBSDiscIndex. Panel
C presents a correlation matrix, with Pearson (Spearman) correlations appearing above (below)
the diagonal. The sample comprises 158 EGC firms.
33
Table 7. Additional Analysis––Index of JOBS Act Provisions and Market Outcomes
Panel A. Underpricing
Variable
EGC
JOBSDiscIndex
Assets
Revenue
BM
ROA
Age
Tech
PctRetained
Big4
Fixed Effects
R2 (%) / N
Underpricing(0)
coef
t-stat
0.56
0.22
1.95
3.11
–1.49
–1.92
2.50
5.23
–3.36
–1.36
3.66
3.75
–1.60
–1.75
0.11
0.07
–2.39
–0.53
1.33
1.02
Industry, Exchange,
Venture, Underwriter
17.19 / 933
Underpricing(0,1)
coef
t-stat
–2.42
–1.03
3.05
5.39
–1.53
–1.92
2.62
6.01
–3.33
–1.20
4.59
6.26
–1.86
–1.83
–1.77
–0.90
–1.11
–0.22
1.32
1.00
Industry, Exchange,
Venture, Underwriter
16.48 / 933
Underpricing(0,30)
coef
t-stat
–9.43
–1.87
6.16
4.56
–2.00
–1.71
2.25
2.56
–1.61
–0.56
6.25
4.73
–0.58
–0.43
1.68
0.50
1.14
0.25
3.72
1.93
Industry, Exchange,
Venture, Underwriter
15.77 / 933
IdiosyncVol
coef
t-stat
–0.52
–1.68
0.21
5.62
–0.20
–4.14
–0.02
–0.49
–0.37
–2.75
–0.07
–0.78
–0.14
–1.90
–0.01
–0.04
0.57
1.63
0.16
1.17
Industry, Exchange,
Venture, Underwriter
27.51 / 933
Beta
Panel B. Return Volatility
Variable
EGC
JOBSDiscIndex
Assets
Revenue
BM
ROA
Age
Tech
PctRetained
Big4
Fixed Effects
R2 (%) / N
Volatility
coef
t-stat
–0.51
–1.23
0.20
4.56
–0.20
–3.85
–0.02
–0.64
–0.46
–3.23
–0.05
–0.59
–0.14
–1.85
–0.06
–0.56
0.59
1.70
0.16
1.20
Industry, Exchange,
Venture, Underwriter
27.97 / 933
coef
t-stat
–0.20
–1.29
0.03
0.98
–0.06
–1.94
0.06
2.07
–0.08
–0.70
–0.05
–0.52
–0.06
–2.07
0.08
0.85
0.03
0.20
–0.03
–0.34
Industry, Exchange,
Venture, Underwriter
6.35 / 933
Panel A (Panel B) of this table presents results from estimating linear regressions of IPO
underpricing (post-IPO return volatility) on EGC, an indicator that equals one for EGC firms,
JOBSDiscIndex, an index calculated as the sum of the number of JOBS Act provisions each firm
applied, and control variables. t-statistics are based on standard errors clustered by industry and
IPO date. See Table 1 for variable definitions.
34
Table 8. Additional Analysis––Individual JOBS Act Provisions and Market Outcomes
Panel A. Underpricing
Variable
EGC
Confidential
ReduceComp
OmitCDA
ReduceAcct
DelayGAAP
Controls
Included
Fixed Effects
R2 (%) / N
Underpricing(0)
coef
t-stat
–1.89
–0.79
6.57
3.96
5.93
3.79
3.13
0.98
–3.66
–1.61
–3.91
–1.25
Underpricing(0,1)
coef
t-stat
–3.19
–1.65
8.73
4.44
8.79
5.97
0.38
0.22
–0.63
–0.27
–4.28
–1.18
Underpricing(0,30)
coef
t-stat
–9.12
–1.95
12.71
3.05
13.51
6.97
4.60
1.16
3.20
1.05
–8.07
–1.96
Yes
Industry, Exchange,
Venture, Underwriter
17.88 / 933
Yes
Industry, Exchange,
Venture, Underwriter
17.17 / 933
Yes
Industry, Exchange,
Venture, Underwriter
16.89 / 933
Volatility
coef
t-stat
–0.56
–4.08
0.31
2.11
0.26
1.31
0.21
1.63
0.25
2.50
0.11
0.64
IdiosyncVol
coef
t-stat
–0.55
–5.17
0.30
1.73
0.30
1.70
0.29
2.48
0.20
2.23
0.06
0.39
Beta
Yes
Industry, Exchange,
Venture, Underwriter
27.83 / 933
Yes
Industry, Exchange,
Venture, Underwriter
27.95 / 933
Panel B. Return Volatility
Variable
EGC
Confidential
ReduceComp
OmitCDA
ReduceAcct
DelayGAAP
Controls
Included
Fixed Effects
R2 (%) / N
coef
0.13
0.31
0.04
–0.41
–0.11
–0.13
t-stat
0.58
2.59
0.31
–4.84
–1.30
–0.83
Yes
Industry, Exchange,
Venture, Underwriter
7.59 / 933
Panel A (Panel B) of this table presents results from estimating linear regressions of IPO
underpricing (post-IPO return volatility) on EGC, an indicator that equals one for EGC firms,
five indicator variables that equal one if the firm applied the respective JOBS Act provision, and
control variables. For parsimony, coefficients on control variables are not reported. t-statistics
are based on standard errors clustered by industry and IPO date. See Table 6 for definitions of
the provision indicator variables and Table 7 for a list of the control variables.
35

Download Report